I have been using Stable Diffusion for a while and I can tell anyone that it certainly requires some skill to get picture which you want. Making models for the AI software is even harder. Anyway, do people want to ban Google for sharing their images? If humans can copy images they saw on the internet, it doesn’t make any sense why software could not. If the pictures are too similar with the original work, then the same rules should apply for both AI and non AI. What I believe is on a thin line, is using authors name in the prompt (either while training model or generating images). Thats similar like stealing recipes or blueprints in my opinion. But mixing thousands of works together is like mixing tons of ingredients. You are basically just telling the AI how much of each ingredient it should add, making your own (or its) recipe.
I don’t consider myself artist. I think the artists are those who trained the models for the AI and not those who are actually using them.
I have been using Stable Diffusion for a while and I can tell anyone that it certainly requires some skill to get picture which you want. Making models for the AI software is even harder. Anyway, do people want to ban Google for sharing their images? If humans can copy images they saw on the internet, it doesn’t make any sense why software could not. If the pictures are too similar with the original work, then the same rules should apply for both AI and non AI. What I believe is on a thin line, is using authors name in the prompt (either while training model or generating images). Thats similar like stealing recipes or blueprints in my opinion. But mixing thousands of works together is like mixing tons of ingredients. You are basically just telling the AI how much of each ingredient it should add, making your own (or its) recipe.
I don’t consider myself artist. I think the artists are those who trained the models for the AI and not those who are actually using them.